Magnets and Cryogenics Division

Superconducting Cyclotrons for Hadron Radiotherapy

J. V. Minervini1, Member, IEEE, A. Radovinsky1, P. C. Michael1, D. Winklehner2, and L. Bromberg1 with thanks to Dr. Eric Forton, IBA for background information on hadrontheray

European Conference on Applied Superconductivity (EUCAS) 17-21 September 2017, Geneva, Switzerland

1Massachusetts Institute of Technology, Science and Fusion Center 2Massachusetts Institute of Technology, Department of Physics, Laboratory for Nuclear Science Cambridge, MA, USA Outline

• What is hadron radiotherapy • Present and future outlook • Cyclotrons for PBRT • Superconducting cyclotrons • Ironless, variable energy, superconducting cyclotron • Summary

J.V. Minervini MIT-PSFC Conventional Radiation Therapy vs. Hadrontherapy

• Most conventional radiation therapy and arc therapy systems use Xrays of a few MeV for cancer treatment • Dose is not delivered to tissues by the photons themselves, but rather through secondary produced by 3 mechanisms

G.F. Knoll – Radiation detection and measurement, Wiley 3 Conventional Radiation Therapy vs. Hadrontherapy

• Results in: ⁃ A decrease of photon numbers following a superimposition of decreasing exponentials ⁃ Some buildup => dose builds-up and then ~exponentially decreases with depth once electron equilibrium is reached

Image courtesy of http://radiologykey.com/radiation-oncology/ 4 Conventional Radiation Therapy vs. Hadrontherapy

• Instead, hadrons lose their energy in matter according to Bethe-Bloch formula. • In short, it results in the famous «bragg peak» dose distribution

5

G.F. Knoll – Radiation detection and measurement, Wiley As a result:

• Hadrons offer the following advantages:  Little radiation before the tumor  No/little radiation at all beyond the tumor  => Lower integral dose per treatment • Leading to potential clinical advantages:  Up to 50% reduced risk of radiation- induced secondary cancer  Drastically lower risk of adverse effects (treatment toxicity, side effects, 6 growth abnormality) – better quality of life PBS – Pencil Beam Scanning

• Advantages:

— Good 3D dose conformity

— “Flexible”

— Low neutron dose

— No need for patient specific aperture • Disadvantages:

— Dynamic system, less safe than passive system

— Layer by layer, slower than scattering

— Lateral penumbra less sharp than with collimation

scattering PBS Examples

Irradiation of surrounding tissues - Hepatocellular carcinoma

Dose to critical Therapy Arc Therapy Tissues (mean Photons dose) Right Kidney 20 Gy 0.1 Gy Lung 12.5 Gy 8.5 Gy

“PBT was found to be a safe and effective local-regional therapy for inoperable HCC. A randomized controlled trial to compare its efficacy to a standard therapy has been initiated” (*)

8

Images Courtesy of Stefan Both, Ph.D -- PENN Radiation Oncology

(*) Bush DA, et al., « The safety and efficacy of high-dose proton beam radiotherapy for hepatocellular carcinoma: a phase 2 prospective trial. » Cancer. 13 (2011) 3053. Irradiation of Surrounding Tissues - Prostate

IMRT Proton Therapy Dose to critical tissues (mean Photons Protons dose) Rectum 20 Gy 6.5Gy Bowel 18 Gy 10 Gy

“Early outcomes with image-guided proton therapy suggest high efficacy and minimal toxicity with only 1.9% Grade 3 GU symptoms and <0.5% Grade 3 GI toxicities” (*)

9

Images Courtesy of Stefan Both, Ph.D - PENN Radiation Oncology

(*) Mendenhall NP, et al. « Early outcomes from three prospective trials of image-guided proton therapy for prostate cancer” Int J Radiat Oncol Biol Phys. 82 (2012) 213. Rhabdomyosarcoma – Side effects

X-Rays Proton Therapy

“Fractionated proton radiotherapy is superior to 3D conformal photon radiation in the treatment of orbital RMS (…) Proton radiation therapy minimizes long-term side effects” (*)

Images Courtesy Torunn I Yock, MD -- Burr Proton Therapy Center Boston USA

(*) Yock, T. et al; « Proton radiotherapy for orbital rhabdomyosarcoma: clinical outcome and a dosimetric comparison with photons. », Int J Radiat Oncol Biol Phys. 63 (2005) 1161. 10 Pediatric medduloblastoma – Side effects

Side Effects Protons Photons Restrictive Lung Disease 3D CRT Proton Therapy 0% 60%

Reduced exercise capability 0% 75%

Abnormal EKGs 0% 31%

Growth abnormality 20% 100%

IQ drop of 10 points at 6 yrs 1.6% 28.5%

Risk of IQ score < 90 15% 25%

“Proton beam therapy has become a standard of care for pediatric cancers… ” (*)

(*) Presentation Dr. Jay S. Loeffler, NPTC/MGH, ASTRO 2001 11 Growing Interest in Proton Therapy Clinical Advantages

Oct. 2015 data from a leading center in the US

12 Increasing Relevance of Proton Therapy.

Cumulative number of PT publications

3000

+13 in 2015 vs 2500 previous year 2000

1500 Publications

1000

500

0

1969 1974 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 2008 2011 2014 Data from https://clinicaltrials.gov/ 1962

Pubmed search with: Proton Therapy or Proton Radiotherapy or Proton Beam Therapy

13 Market Growth: 2016-2021-2030

Worldwide # Rooms/Year 70 CAGR 120 +10% 2016 – 2030 60 52 50 47 43 Rooms 40 40 38 120 34 Per Year 30 28 19 RT Patients 20 6% Treated 10

0 Billions 2014 2015 2016 2017 2018 2019 2020 2021 … 2030 2.5 Order Intake

14 Many Systems Operating

15 Commercial systems at a glance, by geography (2016)

IBA Number of Rooms - US Number of Rooms – EMEA + ROW

HITACHI 18% 22% VARIAN 4% 38% 8% 58% 4% MEVION 48%

PROTOM

Number of Rooms - Japan Number of Rooms – Rest of APAC MELCO

2% SHI 15%13% 21% 31% Commercial systems acc. since 1996 since acc. systems Commercial 9% PRONOVA 6% 7% 41% 55% AVO

16 Typical Systems - Protons

Proteus®PLUS Proteus®ONE

360 m²

1800 m²

17 Carbon and Heavy Ion

18

18 Major components (IBA ProteusONE)

A rotating gantry to set the beam at the Compact super-conducting right angle accelerator for producing the energetic proton beam

Intensity-modulated proton therapy (IMPT) : the most precise form of treatments

Stereoscopic imaging and CBCT at isocentre: Efficient software integration, enabling easy & flexible workflows accurate patient setup, 19 quality images for adaptive treatments Summary: a typical system is composed of

• An accelerator ⁃ Fixed energy (i.e. cyclotrons) => requires energy modulation and selection ⁃ Variable energy (synchrotrons) • A beam transport ⁃ Fixed beam (eye treatment) ⁃ Most usually a gantry able to deliver the beam at various angles ⁃ With PBS as most relevant treatment modality today • A treatment room ⁃ Patient positioning robot ⁃ Rough patient alignment systems (e.g. lasers) ⁃ Imaging systems (Xray, CBCT, CT) • A treatment control room • A therapy safety system, software 20 Accelerators in PT

• (rough) requirements ⁃ Max. energy: 230 (250 MeV) protons – 400 MeV/u carbon ions ⁃ Min energy: ~70 MeV protons ⁃ At least 2 Gy/l/min => a few nA average beam current at nozzle level ⁃ Fast beam intensity modulation . In development or for the (far) future: ⁃ Minimum footprint . Linacs ⁃ Minimum energy consumption . Wakefield accelerators . Cyclinacs • Currently available on the market ⁃ Synchrotrons Beam accelerated on a single path, is ramped =>Variable energy, pulsed beam, multiple-stage ⁃ Cyclotrons and synchro-cyclotrons Acceleration on a spiral path, fixed magnetic field => Usually fixed energy, CW or pulsed (high rep. rate), single stage

21 Synchrotrons

Hitachi “PIMMS” (CERN) design . 70-150 MeV protons . Up to Carbon . Slow cycle . 25 m Diameter . 7 m Diameter . Rep. rate: 5 Hz . Installed @CNAO, Medaustron http://www.protominternational.com/about/about-radiance-330/ Rossi, EurPhysJPlus2011-126-78.pdf

22 Trend is toward more compact machines to reduce costs

Protom . Up to 330 MeV protons . 5 m Diameter, ~16 tons . Being installed @MGH

http://www.protominternational.com/about/about-radiance-330/

23 Challenges and Trends in PT

• Cost & Affordability (multirooms towards 1-2 rooms; cost of a fraction) • IMPT - scanning ⁃ Speed of treatment ⁃ Even faster with continuous line scanning?

• Image guidance and Treatment robustness (leverage on sharp dose distribution) ⁃ Setup errors, Intra-fraction changes, Organ motion management ⁃ Inter-fraction changes, Adaptive treatments  Offline  Online with as little additional imaging dose as possible ⁃ In Vivo & real time range verification • Complexity, and so forth… Sustainability, too

24 Cyclotrons basic principles

• If you neglect relativistic effects, a charged particle in magnetic field orbits at a constant frequency

fp = ω / 2π = q B / 2π m

25 Commercial PBRT Cyclotrons

IBA C230 Varian-Accel Probeam Mevion SC250 IBA S2C2 . 230 MeV protons . 250 MeV protons . 250 MeV protons . MeV protons . 4.3 m Diameter . 3.1 m Diameter . ~1.5 m Diameter . 2.2 m Diameter . CW beam . CW beam (shield) . Rep. rate: 1 . Normal conducting . Superconducting . Superconducting kHz . Magnet: 200 kW (NbTi) . Superconduc (Nb3Sn) . RF: 60 kW . Magnet: 40 kW ting (NbTi) . RF: 115 kW . RF: 11 kW

26 Proton cyclotrons - Ongoing developments

• Isochronous: SHI, Varian/Antaya, Pronova/Ionetix, Heifei/JINR

Ant Derenchuck - NAPAC 2016 Karamysheva - THP20 cyclotrons 2016 Antaya – CAS 2015 SHI Pronova/Ionetix Heifei/JINR Varian/Antaya . 230 MeV protons . 250 MeV protons . 200 MeV protons . 230 MeV protons . 2.8 m Diameter . 2.8 m Diameter . 2.2 m Diameter . 2.2 m Diameter . CW beam . CW beam . CW beam . CW beam . Superconducting . Superconducting . Superconducting . Superconducting (NbTi) (Nb Sn) (Nb3Sn) . 30 tons 3 . 55 tons . 60 tons . 3.6 T (extr.) . 30 tons+ . 4 T (extr.) . 3.7 T (extr) . 5.5 T (extr.) . “Flutter” coils

27 Mevion S250

Cyclotron Weight ~25 t

J.V. Minervini MIT-PSFC Comparison of PBRT Cyclotrons

Mevion S250 IBA S2C2 Varian Proscan IBA C230 R pole (m) 0.34 0.50 0.80 1.05 D Yoke (m) 1.80 2.50 3.10 4.30 Height (m) 1.20 1.50 1.60 2.10

Bo (T) 8.9 5.7 2.4 2.2

Bf (T) 8.2 5.0 3.1 2.9 Mass (tonnes) 25 50 100 250 254 250 235 Tf (MeV) 230/250 Compact Superconducting Cyclotron Research at MIT

• Work started in late 2002

• Initial focus: compact cyclotrons to enable low cost Proton Beam Radiotherapy

• 9T Superconducting Synchrocyclotron was first designed in 2006

• Now we are working on several machines for medical applications, basic science and other advanced applications Compact Superconducting Cyclotrons beyond 6T …

• Compact (a few cubic meters) • Transportable (minimize the mass and power)

• Not tethered to a helium liquefier- Use cryo-coolers; HTS leads; many conductor types • Full acceleration in 1 accelerator stage • High Field Superconducting Cyclotron Possible (>6T)

• E= 10-1000 MeV protons and heavy ions +’s and –’s of Ironless Cyclotrons +’s: - Reduced weight. - Reduced fringe field. - Larger mid-plane and axial bore clear spaces – can use interchangeable (Ion Source/RF/Extraction) cassettes for different Ions (protons, lithium, carbon). - Plenty of space inside the cryostat. - No need to shim the iron – big advantage for mass production. - High factory winding tolerances allow field profile repeatability. - No external iron – no positive magnetic stiffness, simpler cold mass support. - No internal (cold) iron – less load on cryogenics for faster cooldown and warmup. - Scaling laws ease magnetic design process.

-’s: - Somewhat larger radius shielding coils – Increases difficulty of conduction cooling by cryocoolers. - Uses more superconductor than systems with iron - Increases magnet stored energy making quench protection more difficult 32 Further Features of Ironless Cyclotrons

• Scalable beam energy by adjusting coil current– can vary beam energy with extraction at the same radius (restrictions apply) • Shielding coils require small ampere-turns: • Can be LTS • Can be HTS heat sunk to thermal radiation shield • Can be external, copper coils.

33 Ironless k230 Magnet Design

Field Shaping Coils Main Coils

Shielding Coils Coil Sets and Stray Magnetic Field

Main field, field shaping, and field Field magnitude at surface of coils. shielding coils

Stray magnetic fields in the axial cross-section. (10 gauss to 100 gauss with 10-gauss increments) Ironless k230 Magnet Design w/ Variable Energy

Beam B0 T 5.025 Bex T 4.637 Rex m 0.501 Tex MeV 230.0

Coil Em MJ 30.9 Iop A 3,000 L H 6.89 2 Jwp A/mm 52.0 Bmax T 6.61 Magnet Diameter m 3.00 Magnet Height m 2.01 Conductor Mass kg 6,310

Field R(10G) m 3.8 Z(10G) m 3.8

• The magnet Cryostat is an OD= 3m, H= 2m cylinder Magnet and Conductor

• Conductor is NbTi Cable-in-Conduit with sealed supercritical helium  Helium is heat capacity for adiabatic ac loss energy absorption • Conduction cooled by cryocoolers –> Heat removed in time interval between patients.

• Higher Jwp -> lighter

• High-Field Nb3Sn could be used, if desired, for highly compact design -> 4 tons, OD~2 m  Also allows adiabatic absorption of more ac losses. SUMMARY OF MAGNET COMPONENT WEIGHTS

Conductor and Coils NbTi-a NbTi-b All Length (m) m 19289 1052 Mass SC Strand kg 2508 51 2559 Mass Cu Strand kg 1321 0 1321 Mass SS Conduit kg 2722 31 2752 Mass Insulation kg 225 6 231

Total Mass of Coils kg 6551 82 6863 Cryostat Mass Vacuum Vessel kg 3773 Mass Radiation Shield kg 485 Total Cryostat kg 4258 Cold Mass Structure Mass Vacuum Vessel kg 3773 Mass Radiation Shield kg 485 Total Cryostat kg 4258 Cold Mass Structure Total Material SS 316LN Al-6061-T6 2631 Mass kg 1135 1496 Magnet System Total Weight Coils kg 6,863 CM structure kg 2,630 Cryostat kg 4258 Total kg 13,752

More access to the beam area

Iron Free Design Coil Arrangement in a Cryostat Cold Mass Support in a Cryostat

Replaceable RF System Cassettes Field, Beam Energy, and RF Frequency at 230 MeV

B-field vs. Radius Beam Energy vs. Radius

Acceleration Frequency vs. Radius Variable Beam Energy

• Beam energy is varied for range scanning • Change operating current and field in the coils proportionally for variable energy • This requires adjustment of the coil current RF voltage and frequency. • Multiple ramps provide repainting.

• The ranges of variation of these parameters:  1.59 kA < Iop < 3 kA  0.45 kV < V < 0.73 kV  1.16 MW < P < 1.34 MW Variable Beam Energy

• Remain at constant energy while painting a layer during Dtlay=0.5 s increments

• Ramp between layers inDtL2L=0.5 s

• Duration of the cycle is 37 seconds for this scenario of scanning from 230 MeV to 70 MeV • Power and voltage are delivered by the magnet power supply to change the magnet current and beam energy Variable Beam Energy

• Variable beam energy is accomplished by modulating the coil current, which results in linearly proportional change of the magnetic field in the beam space.

• This process has implications both for the magnet itself and for the beam controls.

• For a Tmin=70 MeV and Tmax=230 MeV synchrocyclotron the layer to layer beam energy change was defined as constant

DT = 4.5 MeV linear ramps, each lasting DtLayer-to-Layer = 0.5 seconds.

• The constant beam energy intervals during which the in-

layer painting takes place are set to a constant, Dtlayer=0.5 s, for each layer. Pencil-Beam Scanning (PBS)

J.V. Minervini MIT-PSFC HYSTERESIS LOSSES IN THE SUPERCONDUCTOR PER CYCLE

Conductor NbTi-a NbTi-b

B0 T 0.38 0.38 2 J0 A/m 4.00E+10 4.00E+10

df m 6.50E-06 6.50E-06 alpha 0.53 0.53

Bf T 6.56 5.36 J/m3 5349 5261 E mJ/cm3 5 5

• Hysteresis losses are small. • Adiabatic temperature rise is absorbed by helum in CICC without exceeding current sharing temperature. • This allows for multiple repainting sweeps for a single treatment. • Coils are recooled by cryocooler in between patients. • If layers can be repainted multiple times before energy steps then only a single energy sweep is required. Summary

• Hadron therapy cyclotron accelerators can be improved by replacement of resistive magnets with superconducting magnets. • Ironless cyclotrons are feasible and provide better magnetic shielding. • Superconducting cyclotrons can be much lighter and smaller than conventional systems leading to space and cost savings (physical and operating). • Variable energy synchrocyclotrons are theoretically feasible. Following these engineering studies the step is to build a working prototype. Magnets and Cryogenics Backup Division

J.V. Minervini MIT-PSFC Beam Controls (Beam Energy Variation)

T(r,t)=E0*(sqrt(1+(Kb(t)*B(r,0)*e*r/(m0*c))2)-1)

• Matching T(r,t) can be achieved by a run-time adjustment of the per turn RF voltage. • Quantitative assessment of this variation will be done after defining more parameters of the RF system, per-turn gain, in particular.

Trajectories are Collinear for Scaled Beam Energies T(Rex,t)/T(Rex(0) T gamma Rigidity Bex Kb=B/B0 Jwp MeV T*m T A/mm2 1.000 230.0 1.245 2.322 4.637 1.000 52.0 0.652 202.2 1.215 2.162 4.318 0.931 48.4 0.304 70.0 1.075 1.231 2.459 0.530 27.6

Proton was launched in the circumferential direction from the same spot in the mid-plane at Rex=50.1 cm with the respective energy T(Rex,t), T(Rex,t)/T(Rex(0)=1.0, 0.652 and 0.304

Trajectories of the proton in the axial and lateral perspectives:

• all form perfectly coincident circles

• all lie in the same plane Protons are good - How heavy should we go?

20 MeV alphas . For the same range, heavier and/or more charged particles need higher entrance energy . LET is higher . Straggling is reduced . Biological effect is usually enhanced

5 MeV protons

Ugo Amaldi and Gerhard Kraft - Radiotherapy with beams of carbon ions Reports on Progress in Physics, Volume 68, Number 8 Published 11 July 2005 • 2005 IOP Publishing Ltd How heavy should we go?

. But: . You want to avoid killing upstream cells . Fractionation problem

Ugo Amaldi and Gerhard Kraft - Radiotherapy with beams of carbon ions A. Sessler, Cyclotrons’10 Reports on Progress in Physics, Volume 68, Number 8 http://accelconf.web.cern.ch/Accelconf/Cyclotrons2010/talks/frm1cio02_talk.pdf Published 11 July 2005 • 2005 IOP Publishing Ltd A promising future

• PT has a real potential to grow from ~1% today, up to almost 20% of RT treatments. • The development of referral models is strongly encouraged, and should accelerate the adoption of PT on objective basis.

TODAY PROOF 5-10 YEARS PROOF 15-25 YEARS

Reduced Side Effects (12.1% of RT patients)

Intracranial GI (Esophagus, H&N gastric, rectal, Urologic pancreas) nd Accepted standard Lung Lymphoma Reduced 2 Cancers Sarcoma (0.6% of RT patients) Breast (2% of RT patients) Gynecological Eye Breast Pediatrics Lymphoma Base of skull Testis Brain Improved Local Control (3% of RT patients)

Intracranial H&N National Health Ministry in NL Urologic (prostate & bladder) Independent reports UK and DK Lung (NSCLC) ASTRO Model Policy Sarcoma …

51 Similar magnet design has been built

Similar to MIT designed High Precision Dusty Plasma magnet, currently manufactured for Auburn University by SSI Magnetic Design

Conventional Cyclotron Magnetic Model Iron Free Cyclotron Magnetic Model

• The coil set of the iron free design was optimized to match the magnetic field versus radius, B(r), profile of the conventional design. • The profiles of the proton beam energy vs. radius for both models are almost identical.

Magnetic Field vs. Radius Proton Beam Energy vs. Radius Active Magnetic Shielding

Conventional Cyclotron Magnetic Model Iron Free Cyclotron Magnetic Model

Note different scales

• The lines indicate fields from 10 G to 100 G with 10-G increments. • For the iron free design the 10-G level occurs at a radius less than 2 m, whereas in the conventional design the calculated field at 2 m is 180 G. • In the axial direction the field at the same, 2 m, distance from the iso-center is 10 G and 410 G respectively. Parameter Comparison

Model With Iron Ironless Beam

Bo T 8.877 8.791 Bex T 8.132 8.109 Rex m 0.297 0.296 Tex MeV 247.2 245.7 Coil

Em MJ 9.6 32.0 Iop A 2,000 2,000 2 Jwp A/mm 180.0 180.0 Bmax T 10.98 11.60 Top K 5.0 5.0 dT K 2.5 1.9 OD m 1.80 2.17 H m 1.20 1.61 Mcond kg 1,448 2,225 Field R(10G) m 5.0 1.8 Z(10G) m 8.2 2.0

Weight and Size Comparison

Model With Iron Ironless Parts Density Volume Weight Volume Weight kg/m3 m3 kg m3 kg

Iron Yoke 7,860 2.105 16,545 0 0 Bobbin 7,860 0.299 2,350 0.342 2,686 Windings 8,000 0.181 1,448 0.278 2,225 MLI 24 24

Cold Structure 3,822 4,935

Cryostat 7,860 0.137 1,078 0.184 1,446 Supports 89 65

Thermal Shield 7,860 0.027 216 0.037 289 Cryocoolers 74 74

Magnet 5,278 6,808

Total (Magnet + Iron) 21,823 6,808

J.V. Minervini MIT-PSFC Magnet Design

Beam parameters Maximum central magnetic field (at R=0, Z=0) 4.980 T

Maximum magnetic field at extraction (at R=Rex, Z=0) 4.596 T

Extraction radius, Rex 0.501 m

Maximum beam energy, T(Rex) 226.3 MeV/u Coil Stored magnetic energy, E 31.1 MJ Outer diameter of cryostat, OD 3.00 m Overall height of cryostat, OH 2.02 m Magnitude of fringe magnetic field In radial direction, B(at R=3.5m, Z=0) 11 Gauss In axial direction, B(at R=0, Z=4.5m) 12 Gauss

J.V. Minervini MIT-PSFC